In 2005, we embarked on an RO-1 funded project to build a quantitative multi-parameter protein-based model to predict recurrence in melanoma. The effort culminated in a Melanoma Recurrence Risk (MRR) test that was published in JCO in November 2009. The MRR test was based on 5 markers (two of which required subcellular localization) that could distinguish sentinel node negative patients with a 40% chance of recurrence from those with a 10% chance of recurrence in 8 years. Although this result is quite promising, was validated on 250 sentinel node patients, and has caught the interest of two diagnostic companies, the MMR has not been tested prospectively. Thus we are faced with a choice of "locking" this test and beginning a prospective clinical trial or trying to improve the test, to get better separation between low risk and high risk disease. We have opted to let potential licensees (if there are any) test MMR version 1 while we move on to the development of version 2. Careful review of our work and the melanoma biomarker field has revealed about 75 very promising markers that we have not yet tested for potential inclusion into our MRR assay. Furthermore, our original assay was validated on a single cohort and has not been converted for use on standard melanoma biopsy tissue, since it was constructed and validated on TMAs. Here we propose construction and validation of a second version of the MMR. We will achieve this goal in three aims. The first aim is to construct a new training set and then test a series of new markers on this cohort. These new markers were either lost due to technical issues in our original version, or overlooked in the first round selections. A new model will be constructed based on the new marker data and internally validated. Then in aim 2 we will validate the cohort on two or three independent melanoma cohorts from at least one other institution. In aim 3, working in parallel to the validation in aim 2, we will develop and standardize the MMR version 2 assay for use on whole sections typical of routine melanoma specimens. At the completion of this grant, we anticipate a MMR version 2 assay will be ready for application to patients and for a prospective diagnostic clinical trial. PUBLIC HEALTH RELEVANCE: A key clinical problem in melanoma is the management of patients with melanoma greater than 0.8mm thick and no evidence of tumor in their sentinel nodes. Here we propose to build a model, based on multi-parameter, standardized protein expression in the primary melanoma, that can predict which of these patients have a high likelihood of recurrence and can subsequently be managed more aggressively.